Home 
Contact 
Profile 
Videos 
Written for the 13th Annual SAS Users' Conference held on 15th September 1999 at the JW Marriot Hotel by Prof Dr Omar Hasan Kasule
ABSTRACT
The paper presents experience of SAS over the past 3 years in management and analysis of large epidemiological data sets with up to 2000 records and up to 300 observations per record. In this period 5 major studies were executed. The first study was a questionnaire survey of knowledge, attitudes, and Practice regarding cancer students from many countries. The second study was a hospital-based study of the epidemiology of low birth weight in Kuantan covering 300 mothers. The third was a multi-center nation-wide pilot cancer registration project. The fourth was a nation-side study of knowledge, attitudes, and practice regarding cancer involving a questionnaire pre and post health education intervention. The fifth is an on-going longitudinal study of the impact of the medical curriculum on the knowledge, attitudes, and practice of medical students regarding cancer. All the studies involved more than 200 variables. Overall there is satisfaction with the SAS software. The author first used SAS in the US in 1982 and has noticed that over the past 16 years it has come a long way in terms of being user friendly and requiring hardly any programming ability. The interface with microsoft and excel has been a very attractive feature. The analyses carried out were mostly the one-way or 2-way descriptive analysis. Logistic regression was used extensively in multivariate analyses. The handling of large data sets is efficient. We have learned the importance of using numerical variables as much as possible to avoid future problems. We have not yet used the SAS graphics packages but plan to do so soon. It is recommended that SAS be made more user-friendly by incorporating features that allow the analyst to insert denominator data for incidence rate computation.